ES2949803T3 - Low gain avalanche detector for low penetration particles - Google Patents

Abstract

The low-gain avalanche and low-penetration particle detector comprises a multilayer structure and receives particles from a radiation source (13). It consists of a thin entrance region that receives the particles from the radiation source (13); a low penetration particle detection region, with a surface field stop p++ (1), placed below the entrance region, and an absorption layer p (3), placed below the surface field stop p++ (1) , and an n(4) multiplication layer; and a high-penetration particle sensing region placed beneath the n-multiplying layer (4), which consists of an n-silicon substrate (5). Due to the chosen doping polarities, the primary electrons (created by the radiation source particles (13)) move away from the entrance zone. In this way, signals from low-penetration particles or radiation experience amplification. (Automatic translation with Google Translate, without legal value)

Description

DESCRIPTION

Low gain avalanche detector for low penetration particles

Object of the invention

The object of the invention is a low-gain avalanche detector for low-penetration particles with a stratified structure, which has the ability to detect radiation or low-penetration particles, excellent temporal resolution and a low noise level, and which also It is capable of operating at room temperature.

Background of the invention

Silicon radiation detectors, based on the PIN diode, are well established in high-energy physics experiments. The PIN diode, operated with an external reverse bias at full depletion, produces a signal proportional to the energy deposited by the incident radiation.

Silicon photodetectors are widely used, for example, in synchrotron facilities as X-ray detectors or in large-scale particle physics experiments as particle detectors. Silicon photodetectors are used for their outstanding features, which include their compact size, high quantum efficiency over a wide wavelength range, and high spatial resolution, defined by diode segmentation as either plates, strips, or pixels.

To detect low-penetration particles (e.g., low-energy protons, electrons, and/or neutrons) and low-energy X-rays (light particles, photons), the avalanche photodiode (APD) was developed. This type of detector features internal signal amplification with a gain proportional to the applied bias voltage, with good uniformity over large detection areas. The operation of such a device in the linear region provides a signal output proportional to the absorbed energy, with a gain factor between 10 and 100, allowing the detection of low intensity signals, even up to a few photons. However, there is noise associated with the multiplication process, which can distort the signal-to-noise ratio, and the leakage current levels associated with APDs are too high to be compatible with modern hybrid pixel readout systems being used. used in particle physics and hybrid pixel detectors.

Modifying the doping levels in the APD allowed the production of a device with a lower gain, in the region of 5 to 10, operating in the linear mode for a proportional response. Such devices are known as low gain avalanche detectors (LGAD).

These operate at low voltage and therefore require a high resistivity silicon substrate (5-10 kücm) to achieve full depletion. The lower gain reduces noise and the detector's gain dependence on device temperature and applied bias voltage, compared to standard APDs. The diode can be segmented in the same way as standard PIN detectors, as the low gain prevents crosstalk. Therefore, hybrid pixel detectors can be fabricated that allow detection of low-energy X-rays and produce signals below the noise floor of the electronics.

The LGAD design enables the detection of subnanosecond signals produced by minimal ionizing particles, where ionization occurs uniformly as a function of depth in the detector. For an LGAD with a gain of 10, if the LGAD substrate is reduced by an order of magnitude compared to a standard PIN silicon detector, the same total signal loading will be produced from a minimal ionizing particle. However, the collection time will be reduced by an order of magnitude.

Additionally, inverted low gain avalanche detectors (iLGAD) are known from the state of the art. iLGADs have an APD (avalanche photodiode) structure at the rear contact, while the segmented front surface is created with an ohmic contact. iLGADs are P-on-P position sensitive detectors, with low signal amplification and a uniform electric field across the device area, ensuring the same signal amplification wherever a particle passes through the sensitive bulk.

The technology is repeatable and well controlled. However, it is not capable of detecting ions, charged particles or radiation with shallow penetration depth, that is, penetration depths less than one micrometer.

The document "Low Gain Avalanche Detectors for high energy physics", P. Fernández-Martínez et al., 2015, X Spanish Conference on Device Electronics (CDE), IEEE, 02/11/2015, discloses an LGAD of high penetration with an n+ electrode, a p-type multiplication layer and a p-substrate, where the particles are absorbed into the substrate and where the corresponding generated carriers (electrons) are amplified in the high-field multiplication layer.

Description of the invention

A low penetration particle is understood to be a particle with a depth of penetration into the silicon of less than 1 micrometer. When trying to detect the signal of a low-penetration particle, a sensor faces three distinct requirements: first, the non-active layers on the surface must be thin, second, the charge collection efficiency (CCE) must be high at the surface and, thirdly, the reading must have a high signal-to-noise ratio (SNR).

Regarding the first requirement, semiconductor sensors always have non-active layers on their surfaces to protect them from the external environment. These non-active layers should be made as thin as possible to allow the low-penetration particle to reach the sensitive layer of the sensor in the first place. But even so, a significant part of the energy of the primary particle is deposited in the non-active layers and cannot be measured. Therefore, the thickness and composition of the non-active layers define the lower energy limit of the sensor.

Regarding the second requirement, the surface of a silicon substrate (below the non-active layers) has a higher defect density than the bulk and is typically heavily doped to form a field stop. Therefore, the electron-hole pairs created from the energy deposited in the silicon surface region face a higher probability of recombination, thus further reducing the measurable signal. This manifests itself in a CCE that is typically significantly less than 1 near the silicon surface. However, a sensor with high detection efficiency should have a CCE of 1 near the surface.

As for the third requirement, the charge of the remaining signal can easily be an order of magnitude lower than the signal of a minimally ionizing particle (MIP). Thus, a reading with a sufficiently high SNR is of utmost importance.

Especially for low-penetration particles, such as low-energy protons, the sensor signal is created near the sensor surface, mostly at a depth less than 1 micrometer. In conventional silicon sensors, the non-active layers on the surface are much thicker than that.

The proposed low-gain avalanche detector for low-penetration particles (pLGAD) adapts to these special requirements of low-penetration particles by comprising a thin entrance region, on the order of 15 nanometers, preferably made of aluminum, aluminum oxide or silicon nitride. The polarity of a signal collecting electrode is n-type, so that the primary electrons move towards the reading region by crossing a multiplication layer. This also means reversing the polarity of the silicon substrate to maintain the p-n junction and thus the multiplication layer in the input region. The p-n junction and multiplication layer are moved to the bulk, away from the input region, and a thin p-type absorption layer is introduced.

Therefore, the Low Penetration Particle Low Gain Avalanche Detector (pLGAD) of the present invention enables the detection of ions, charged particles or radiation with low penetration depth (i.e., less than one micrometer). Specifically, it comprises a multilayer structure that combines an ultra-shallow input region with a linear multiplication layer and low capacitance, with the possibility of fine segmentation (few micrometers) of the electrodes in the reading region.

In the pLGAD detector, only signals created near the input region are amplified. This makes this detector concept less interesting for high-energy particle applications, but perfectly suitable for low-energy physics experiments and respective applications in the field of medical and materials science, testing or diagnosis. Note that in this pLGAD detector concept, most of the detector's intrinsic leakage current remains unamplified, as this current is created in the bulk of the material, and only the gaps cross the multiplication layer. The pLGAD detector concept therefore makes clever use of two systematic asymmetries to detect low-penetration particles: on the one hand, the signal is created only near the surface (input region) and, on the other hand, the fact that that only the electrons multiply.

This detector can be applied in fields such as neutron physics, ion physics, medical physics, space research, synchrotron detectors and quantum technology, for the detection of protons, electrons, neutrons and/or low-energy X-rays.

Specifically, the detector comprises a multilayer structure with five main regions: an entrance region, a low-penetration particle detection region (p-doped region), a high-penetration particle detection region (n-doped region), a reading and a peripheral region.

The entrance region faces a radiation source, receiving a variety of particles. It may comprise a protective coating and a p++ surface field stop doping in front of the radiation source or located below the protective coating, tens of nanometers thick.

The low-penetration particle detection region is located below the entrance region and comprises an absorption layer p, located below the surface field stop p++, and a multiplication layer n, located below the absorption layer p , approximately 1 to 5 micrometers thick.

The high penetration particle detection region is located directly below the low penetration particle detection region, still close to the entrance region, and comprises an n-- silicon substrate, after the n multiplication layer, approximately 10 to 1000 micrometers thick.

The device also comprises a readout region, which contacts the high penetration particle detection region, which may comprise plate, pixel or strip electrodes. The electrodes comprise, in turn, n++ contacts, in contact with the n-- silicon substrate, and metallic contacts, connected to the n++ contacts.

The readout region may additionally comprise p+ stops, which alternate with the strip electrodes and which are in contact with the n-- silicon substrate.

Finally, the device also comprises a peripheral region, which in turn may comprise one or more metal contacts attached to the protective coating; one or more metallic contacts, linked to the p++ surface field stop doping, which allow the p++ surface field stop doping to be placed at a defined electrical potential; a silicon oxide layer, located under the protective coating; and a p+ termination doping, located laterally on either side of the p++ surface field stop. The material polarity of pLGAD makes the primary electrons,

created by the incoming particle, move towards the reading region. Thus, the multiplication acts only on the signals created in the detection region of low penetration particles.

Due to its design and its application in low penetration particles, pLGAD has many interesting advantageous properties:

- noise performance and background rejection can be optimized by appropriately choosing the detector thickness.

- The SNR of low penetration particles increases without problems.

- Current pulse shape allows for fast timing.

- Operation at room temperature is feasible.

- Flexibility in choosing the energy range of incoming particles for full amplification. - The multiplication layer supports many reading schemes.

The following paragraphs detail the advantages listed:

Noise performance and background rejection can be optimized by appropriately choosing the detector thickness. The main signal is created near the surface. Therefore, unlike high-energy particles, the total charge depends only on the energy of the incoming particle and not on the thickness of the detector. This means that the thickness of a pLGAD can be freely chosen within the limits of production technologies, to optimize, for example, noise behavior or response to background particles.

The thickness influences the duration and amplitude of the pulse and the charge capacitance observed by the readout electronics. Additionally, a thin pLGAD (below 100 micrometers) would allow the passage of high-energy particles and suppress their signal height, and a thick pLGAD (above 1.5 millimeters) would allow deep-penetrating particle spectroscopic measurements. stopping them completely inside the sensor.

The SNR increases smoothly. With the choice of pLGAD polarities, only the ionization electrons (primary and secondary electrons) created by the incoming low-penetration particles are multiplied. Electrons created in the bulk, such as from a moving background particle or by thermal excitation, move away from the multiplication layer. This means that, with moving particles, a pLGAD behaves like a planar sensor.

The leakage current and its corresponding noise are not amplified. This means that the SNR for low-penetration particles increases smoothly through the gain, unlike a flat sensor.

The current pulse shape allows for fast timing. The primary electrons pass through the thickness of pLGAD in a straight path. Unlike competing technologies, in a pLGAD sensor, the primary electrons pass through the thickness of the detector in the shortest possible path. Likewise, only electrons cross the detector, and not holes. This results in a short pulse duration with a fast rising edge, which is beneficial for timing.

Operation at room temperature is feasible. High detection efficiencies can be achieved with detectors that store the signal charge and read it repeatedly. This, however, is a time-consuming process, during which the leakage current distorts the signal. Consequently, cooling is essential for these sensors. In contrast, a pLGAD can operate at room temperature, while achieving high detection efficiencies.

The energy range of the incoming particle can be adjusted for full amplification. The composition of the entrance region defines the minimum energy for a low-penetration particle to be detectable. In contrast, the depth at which multiplication occurs defines the higher energy an incoming particle can have, while its entire signal is multiplied. After all, to be fully amplified, an incoming particle must come to a complete stop in the p-absorption layer.

The multiplication layer supports many reading schemes. The creation and reading of the signal occur on different faces of the detector. Therefore, for a planar reading electrode, a pLGAD can be made with different geometries, for example, in the form of a silicon microstrip sensor or a pixel arrangement, or any other reading scheme that can be implemented with n-type electrodes in an n-type substrate. This includes, among others, readout schemes that integrate the first amplification phase into the detector, such as HV-CMOS sensors.

If a pLGAD is made with a flat reading electrode, it is as easy to operate as other flat sensors. Or, put another way, in principle, a pLGAD-like multiplication layer can be added to existing sensor technologies to improve their detection efficiency for low-penetration particles.

Description of the drawings

To complement the description made and in order to contribute to a better understanding of the characteristics of the invention, according to a preferred example of a practical embodiment thereof, a set of drawings is attached as an integral part of said description. , where, for illustrative and non-limiting purposes, the following has been represented:

Figure 1- Shows a first embodiment of the low-gain avalanche detector for low-penetration particles, in a cross section of half a detector.

Figure 2 - Shows a second embodiment of the low gain avalanche detector for low penetration particles receiving low penetration radiation in a cross section.

Figure 3 - Shows a simplified cross section of the device with the electric field profile.

Figure 4 - Shows the comparison of the current pulses between a 15 keV proton (top) and a minimally ionizing particle (MIP, bottom), both with normal incidence of the projectile particle.

Figure 5 - Shows the current pulse created by a 15 keV proton as a function of time, detailed in the first peak of figure 4.

Figure 6 - Shows the undetected protons with different energies at normal incidence, depending on the detection threshold of the data acquisition system in primary electrons.

Preferred embodiment of the invention

With the help of Figures 1-6, the preferred embodiment of the present invention is described below.

As shown in Figure 1, the low-gain avalanche detector for low-penetration particles comprises a multi-layer structure with:

- an input region, comprising:

or a protective layer (6),

or a surface field stop p++ (1), located under the protective layer (6) with tens of nanometers thickness,

- a low penetration particle detection region, comprising:

or an absorption layer p (3), located below the surface field stop p++ (1),

or a multiplication layer (4), located below the absorption layer p (3), approximately 1 to 5 micrometers thick,

- a high penetration particle detection region, comprising:

or an n--(5) silicon substrate, after the n-(4) multiplication layer, approximately 10 to 1000 micrometers thick,

- a reading region, comprising:

or pixel or strip electrodes, after the high penetration particle detection layer, comprising:

■ n++ contacts (10), in contact with the n--(5) silicon substrate,

■ additional metal contacts (12), connected to the n++ contacts (10),

or p+ stops (11), which alternate with the strip electrodes and which are in contact with the substrate of silicon n--(5), and

- a peripheral region, which includes:

or one or more n++ ohmic contacts (9), located next to the p absorption layer (3),

or metallic contacts (7a, 7b), wherein a first metallic contact (7a) is attached to the surface field stop p++ (1) and one or more second metallic contacts (7b) are attached to the ohmic contacts n++ (9) ,

or a layer of silicon oxide (8), located below the protective layer (6), and

or a p+ (2) termination doping, situated laterally on either side of the p++ (1) surface field stop.

As shown in Figure 2, low energy protons (or other low penetrating particles or radiation) from a radiation source (13) illuminate the detector in the entrance region. The protons will pass through the protective layer (6), which has been placed, for example, by atomic layer deposition to obtain, for example, Al, AI ² O ³ , or by low pressure chemical vapor deposition methods to obtain, for example, YES ³ N ⁴ .

The protective layer (6) is used to protect the surface of the detector from moisture or scratches and ensures long-term electrical stability when operating at high bias voltages. The protective layer (6) can act as an anti-reflective layer for visible, ultraviolet and infrared radiation. The protective layer (6) can also be used as a conversion layer for incoming neutrons. The minimum deposited thickness is on the order of a few nanometers and can be optimized for different radiation wavelengths.

The p+ (2) termination doping of the peripheral region should be formed laterally next to the p++ surface field stop, in order to avoid electric field peaks and ensure stable operation at high voltage.

The incoming protons will also have to pass through the highly doped region, the p++(i) surface field stop, and must therefore be kept as thin as possible to avoid recombination of the electron-holes generated by them. The p(3) absorption layer can be created, for example, by implantation of ions and a variety of dopants (B, Ga, Al).

Specifically, the depth of the p++(1) surface field stop must be less than 100 nanometers, since low-energy proton particles will penetrate on the order of several hundred nanometers. Therefore, the p(3) absorption layer of, say, 1 micrometer thick is placed underneath to absorb all the energy deposited by the proton (or other low-penetration particles). To obtain a multiplied signal of the accumulated charge, the low-penetration particle detection section and the n-type multiplication layer are placed above the high-penetration particle detection layer, resulting in a final triple-doping profile. .

Thus, at the p/n junction a high electric field is created, that is, greater than > 3.105 V/cm, shown in figure 3, where the primary electrons of the electron-hole pairs created will multiply ( e-/ h ⁺ ).

The p absorption layer (3) and the n multiplication layer (4) can be created by ion implantation or by epitaxial growth. Control of doping profiles and forms is important to achieve adequate electron gain and avoid premature breakdowns.

The thickness of the p(3) absorption layer versus the n(4) multiplication layer, which form the p-n junction, can be tailored to match the penetration depth of incoming particles or radiation.

The n--(5) silicon substrate decreases capacitance and improves stability with respect to changes in bias voltage, which directly relates to capacitance. The thickness of the high penetration particle detection layer is designed to reduce input capacitance to the readout electronics and therefore noise. The possible values of its thickness will range between 10 micrometers and 1000 micrometers (or even higher, if technology allows). The thickness of the high-penetration particle detection layer defines the minimum operating voltage of the detector, since to obtain the minimum capacitance, the entire substrate must be depleted of charges.

The reading region of the detector, which may be at its rear, is segmented with n++ contacts (10) connected to the reading electronics through additional metal contacts (12). The n++ (10) contacts can be formed, as shown in Figure 2, with phosphorus or equivalent dopants by ion implantation and with different geometries, such as plates, pixels or strips of different sizes. In principle, any readout scheme compatible with n-in-n contacts can be used, although it depends on the concentration of the multiplication layer n (4) and the absorption region p (3). That includes, but is not limited to, monolithic pixels like those used in HV-CMOS sensors.

Preliminary simulation results show that the avalanche process refers to the doping concentration of the p(3) absorption region and the n(4) multiplication region of the proposed structure. In this sense, it is possible to tune the amplitude of the maximum electric field to adjust its electrical performance, optimize the gain against the anode voltage to increase its linear evolution before the breakdown voltage.

In a planar sensor, the measurable signal is a current pulse created by the movement of charge carriers toward the electrodes. The current pulse expected in the detector of the invention must be simulated and the current pulses of the primary electrons, primary holes, secondary electrons, secondary holes and summed current pulse must be divided.

Figure 4 shows the superposition of these current pulses created by a 15 keV proton (top, low penetration) and an MIP (bottom, passing through the entire thickness of the sensor), both at normal incidence. Only signals close to the input region are amplified. For a very energetic moving particle, for example a MIP, a pLGAD behaves like a classic planar sensor.

Figure 5 shows in detail the current pulse created by a 15 keV proton. The most important characteristics of the pulse shape are (chronologically ordered):

- Point 1 (1 ps): a small drop when the primary holes are collected in the entry region (loose dotted line).

- Point 2 (4 ps): a sharp increase when primary electrons multiply and secondary electrons are created (thick dashed line).

- Point 3 (11 ps): a small drop when the secondary holes are collected in the entry region (thick dotted line).

- Point 4 (26 ps): a downward slope as the secondary electrons leave the highly doped multiplication layer (thick dashed line). This marks the end of the initial spike.

- Point 5 (30 ps): a plateau when the primary and secondary electrons move toward the readout region in a compact package (loose dashed and thick dashed line).

- Point 6 (3 ns, shown in Figure 4): a slight increase when the primary and secondary electrons reach the surrounding reading electrode and experience a higher weighting field.

- Point 7 (3.7 ns, shown in Figure 4): a sharp drop when the ionization electrons (primary and secondary electrons) are collected at the readout electrode.

The exact shape of the current pulse depends on many things, such as the doping profiles, the depth of the p-n junction, and the thickness of the multiplication layer. The current pulse shown in Figure 5 is the result of a simplified simulation setup; realistic doping profiles will result in a more faded shape. However, the main characteristics of the pulse shape will still be present.

The pLGAD concept targets a specific niche: ultra-low noise, position-resolving silicon sensor, with thin entrance region for backlighting. This niche is also shared by two other silicon technologies: silicon drift detector (SDD) and depleted P-channel field effect transistor (DEPFET).

SDDs are a common choice in low-energy physics, especially where large areas must be instrumented and only coarse position resolution is required. Comparing them to a pLGAD would not be fair, as an SDD shines with large area detection elements on the order of square centimeters. This size is much larger than the target position resolution of the pLGAD sensors, so we do not have noise estimates directly comparable to the SDD. Consequently, the SDD is mentioned for completeness, but a detailed comparison of the respective performances is not presented.

DEPFET sensors have their origins in high-energy particle research and are available with small pixel sizes, down to tens of micrometers. In combination with drift rings (as in an SDD), larger pixel sizes are possible.

We can directly compare the performance of pLGAD and DEPFET in Figure 6 at a pixel size of 0.4 x 1.6 square millimeters. A DEPFET sensor with the appropriate pixel geometry achieves a detection threshold below 30 primary electrons, if cooled to -50°C. pLGAD is limited to more than 50 primary electrons, but at room temperature.

Both DEPFET and SDD technologies present a number of complications: the sensors are complex and therefore expensive, and require at least 7 (5 for SDD) different operating voltages. For a DEPFE ^t sensor, cooling is mandatory to keep the leakage current and its corresponding noise low and therefore increase the detection efficiency (see Figure 6). Additionally, DEPFET technology involves several different custom readout chips and has a relatively low frame rate on the order of microseconds.

In contrast, the proposed pLGAD can operate on a single supply voltage, works with commercial readout systems that can read planar sensors, can be operated at room temperature, has high intrinsic time resolution, and is much cheaper thanks to the process. simple planar production.

Figure 6 shows the percentage of undetected protons as a function of the detection threshold of the data acquisition system, for different proton energies at normal incidence, and with a protective layer (6) of 15 nm.

For clarity, Figure 6 does not have error bars, but due to the statistical variation of the Monte Carlo method, the values represented may be higher or lower. Protons can go undetected due to backscatter, when they deposit too little energy in the active region of the detector, or when they stop completely in the non-active layer.

With proton energies of 15 keV, the pLGAD outperforms uncooled DEPFET sensors with the same pixel design and achieves almost the same detection efficiency as a cooled one. In the best case, a proton loss on the order of 0.15% (or a detection efficiency of 99.85%) seems possible.

Claims (11)

1. A low gain avalanche detector for low penetration particles with a multilayer structure, configured to receive particles from a radiation source (13), wherein the detector comprises: - an entrance region, configured to receive the particles from the radiation source (13), and comprising: or a shallow field stop p++ (1), and - a low penetration particle detection region, comprising the following layers, in increasing depth: or an absorption layer p (3), located below the surface field stop p++ (1), or a multiplication layer n (4), located below the absorption layer p (3), - a high penetration particle detection region, comprising: or an n--(5) silicon substrate, after the n-multiplying layer (4), below the low-penetration particle detection region, - a reading region, which makes contact with the high penetration particle detection region, and - a peripheral region, comprising: or a p+ (2) termination doping, located laterally on either side of the p++ (1) surface field stop, and or one or more metal contacts (7a), attached to the surface field stop p++ (1). 2. The detector according to claim 1, wherein the input region additionally comprises a protective layer (6), located over the p++ surface field stop (1). 3. The detector according to claim 2, wherein the protective layer (6) is a conversion layer for incoming neutrons. 4. The detector according to claim 2, wherein the peripheral region additionally comprises a silicon oxide layer (8), located below the protective layer (6). 5. The detector according to claim 1, wherein the surface field stop p++ (1) is within the range of tens of nanometers in thickness. 6. The detector according to claim 1, wherein the multiplication layer n (4) is within the range of 1 to 5 micrometers in thickness. 7. The detector according to claim 1, wherein the n--(5) silicon substrate is within the range of 10 to 1000 micrometers in thickness. 8. The detector according to claim 1, wherein the peripheral region further comprises n++ ohmic contacts (9), located on both sides of the absorption layer P (3). 9. The detector according to claim 1, wherein the reading region further comprises segmented electrodes, after the high penetration particle detection layer, opposite the low penetration particle detection layer. 10. The detector according to claim 9, wherein the segmented electrodes comprise n++ contacts (10), in contact with the n--silicon substrate (5), and additional metal contacts (12), connected to the n++ contacts (10 ). 11. The detector according to claim 9, wherein the reading region further comprises p+ stops (11), which alternate with the segmented electrodes and in contact with the n--silicon substrate (5).